Have you ever felt you waited too long at a traffic light? Well, it seems your wait may have been part of a heavily calculated and intricately craft risk assessment profile run by your city and low profile data crunching firms.
DataKind just happens to be one of those low profile data crunching firms using next level data science to help cities plan out the complexities of reduced driving visibility in certain areas, speed zones and much more.
However, DataKind doesn’t always work alone, the non-profit company recently partnered with Microsoft to tackle three of America’s highly traffic congested cities such as New York, Seattle, and New Orleans.
In Seattle, the city focused on bicycle and pedestrian safety issues. The team’s models identified collision patterns and factors that contributed to higher levels of injury severity, including whether a motor vehicle is making a right turn or left turn, the variation in time of the year and the effectiveness of crosswalks in reducing crash severity. They also identified key variables affecting the likelihood of accidents taking place on particular stretches of road, including traffic volume, land use, number of traffic lanes, street width and pedestrian concentration. Seattle recently passed a levy to fund multi-modal transportation improvements citywide. The results from this project, along with additional safety studies, will help guide more than $300 million in Vision Zero investments over the next nine years.
In New York and New Orleans, DataKind was able to create and offer a new exposure model capability that should help the city of New York avoid injuries and fatalities as well as a new Impact Assessment tool for New Orleans that compare locations for advanced street treatment.
While Microsoft is no stranger to combining data science and civic programs, the partnership with DataKind demonstrates a concerted effort to give machine learning real and life-saving applications.
As Microsoft typically does, it has made the case study available for anyone interested in learning about the various safety models used by the company and DataKind.